JitOperator#

public enum JitOperator#

The subset of libcudf row-IR operators exposed through the Java AST API.

Standard comparison, logical, and mathematical operations remain available through BinaryOperation and UnaryOperation. Operators ending in _OVERFLOW and CHECK_PRECISION are fallible and support JitErrorPolicy.NULLIFY. Other operators require the default JitErrorPolicy.PROPAGATE policy.

NOTE: This must be kept in sync with jni_to_jit_operator in CompiledExpression.cpp!

Enum Constant Summary

`ABS`
`ABS_OVERFLOW`
`ADD`
`ADD_OVERFLOW`
`BITWISE_SHIFT_LEFT`
`BITWISE_SHIFT_RIGHT`
`CAST_TO_BOOL8`
`CAST_TO_DECIMAL128`
`CAST_TO_DECIMAL32`
`CAST_TO_DECIMAL64`
`CAST_TO_FLOAT32`
`CAST_TO_FLOAT64`
`CAST_TO_INT16`
`CAST_TO_INT32`
`CAST_TO_INT64`
`CAST_TO_INT8`
`CAST_TO_UINT16`
`CAST_TO_UINT32`
`CAST_TO_UINT64`
`CAST_TO_UINT8`
`CHECK_PRECISION`	Verify that decimal input 0 fits the INT32 precision supplied by input 1.
`COALESCE`	Return the first non-null input.
`DIV`	Divide without reporting arithmetic errors.
`DIV_OVERFLOW`	Divide and report division-by-zero and signed-overflow errors.
`IF_ELSE`	Select input 0 when input 2 is true, otherwise input 1.
`MOD`
`MOD_OVERFLOW`
`MUL`
`MUL_OVERFLOW`
`NEG`
`NEG_OVERFLOW`
`PREDICATE`	Convert a nullable boolean input into an always-valid predicate.
`RESCALE`	Rescale a decimal input to the target scale supplied to JitOperation.
`SUB`
`SUB_OVERFLOW`

Enum Constants

PREDICATE#: Convert a nullable boolean input into an always-valid predicate.

DIV_OVERFLOW#: Divide and report division-by-zero and signed-overflow errors.

CHECK_PRECISION#: Verify that decimal input 0 fits the INT32 precision supplied by input 1.

RESCALE#: Rescale a decimal input to the target scale supplied to JitOperation.